Aritificial intelligence

With weak AI, the computers have a specialty, which means they are really good in one area. Typical examples are the classic chess computer or Siri. With strong AI, on the other hand, the system has the same intellectual abilities as the human brain. But how does a weak AI become a strong AI?
The first step is to imitate how the human brain learns: an artificial neural network must be created. The artificial neurons are connected to each other and now have to learn. Let’s take the example of image recognition at this point: In the first step, the system is fed with a huge data set of images. Only then does the real work begin: the computer must now recognize which of the images are the “right” ones and receives feedback for each decision. Depending on the feedback, the network changes the connections between the individual neurons. Those that have led to the right result are reinforced. Gradually, the network becomes an intelligent neural network that optimizes itself in advance. Experts call this process deep learning. Is that realistic?

In fact, Shell is already using AI to increase safety at its filling stations. By using cameras on site, algorithms can be set up so that machines learn to recognize situations with increased risk: When customers smoke while refueling, theft or negligent driving.
Microsoft developed the pioneering InnerEye project, which is currently being used at Addenbrooke’s Hospital in Cambridge. Here, machine learning and computer vision are used to analyze and evaluate radiological images. InnerEye makes it possible to detect tumors in seconds and thus make better diagnoses and initiate more efficient treatment therapies.

What makes artificial intelligence so much more attractive than human intelligence?

The cost factor

Machines are much cheaper than employees. You don’t have to pay taxes or duties for them and they don’t take sick leave.

The speed of execution: while a doctor needs around 10 minutes to make a reliable diagnosis, AI can make a million in the same time.

The operational capabilities: Machines show no downtime in their work due to saturation and they don’t need vacation, or rest.minutes it takes to make a certain diagnosis, AI can do a million in the same time.

Accuracy: Machines are more precise, make better decisions and act without the influence of emotions.

There are natural limits to the productivity of the human brain. Even experts repeatedly reach the limits of what is possible. There are no experts who are capable of efficiently processing thousands of pieces of data. The use of AI means that problems like this can be solved completely.